An introduction to statistical energy analysis of structural vibration

Abstract The ideas of the approach to vibration analysis called Statistical Energy Analysis (SEA) are explored without going into great technical detail. The aim of this description is to give guidance to those with particular vibration problems who may ask whether they should be using SEA and, if so, what expectations they should have of it. In the first section, SEA in its most common form is illustrated by the simplest example. In the second section, the question of the underlying assumptions of SEA is considered by a simple and apparently novel approach. This discussion also gives some information on possible methods of measuring the SEA parameters in a given problem and deciding whether a SEA model is indeed appropriate for that problem. We also attempt to give guidance on how SEA should be applied to a given problem, especially how the system under study should be divided up into subsystems. One area in which SEA can be especially useful is in the design of a systematic sequence of experiments on full or model scale when trying to pin down the source of a particular vibration problem: SEA can provide a succession of models, starting with the simplest possible and becoming progressively more complicated, which enable results to be interpreted and suitable questions to be asked for the next stage of testing. This is a more valuable service than is commonly appreciated: without such guidance, an experimenter faced with a complicated structure can spend a very long time making measurements which turn out not to address the problem in hand.